Re: Requirements Document Updated

Alex Milowski wrote:
> Henry S. Thompson wrote:
>> With respect, I think it's important _not_ to conflate inputs and
>> parameters.  This connects up with our discussion on the call
>> yesterday:  inputs, in the prototypical case at any rate, are XML
>> documents, whereas parameters are (or are adequately modelled as)
>> name-value pairs.  Not at all the same thing, and confusing to try to
>> use a single term to cover both.
> Here I have to agree as I view parameters as something that might be
> a simple value (like a string) or an XML document, while inputs
> are *only* XML documents.
> It would, of course, be really nice that it is easy to hook up an
> output of a step to a parameter of another step.

I am not 100% we have built a strong case for parameters. It looks like 
they are only needed for XSLT. Do we have other use cases?

This was one of the motivations to go toward the XDM: whether you are 
talking about "inputs" or "parameters", there would be no difference 
with the XDM.

My fear is that now, will will have two concepts: "input" and 
"parameters". They won't work the same, and the question of how or 
whether a step can produce data which gets connected to a "parameter" 
remains open.


Received on Friday, 10 February 2006 17:05:46 UTC